ArticlePDF Available

COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning: A social network analysis

Authors:

Abstract and Figures

The COVID-19 outbreak brought online learning to the forefront of education. Scholars have conducted many studies on online learning during the pandemic, but only a few have performed quantitative comparative analyses of students’ online learning behavior before and after the outbreak. We collected review data from China’s massive open online course platform called icourse.163 and performed social network analysis on 15 courses to explore courses’ interaction characteristics before, during, and after the COVID-19 pan-demic. Specifically, we focused on the following aspects: (1) variations in the scale of online learning amid COVID-19; (2a) the characteristics of online learning interaction during the pandemic; (2b) the characteristics of online learning interaction after the pandemic; and (3) differences in the interaction characteristics of social science courses and natural science courses. Results revealed that only a small number of courses witnessed an uptick in online interaction, suggesting that the pandemic’s role in promoting the scale of courses was not significant. During the pandemic, online learning interaction became more frequent among course network members whose interaction scale increased. After the pandemic, although the scale of interaction declined, online learning interaction became more effective. The scale and level of interaction in Electrodynamics (a natural science course) and Economics (a social science course) both rose during the pan-demic. However, long after the pandemic, the Economics course sustained online interaction whereas interaction in the Electrodynamics course steadily declined. This discrepancy could be due to the unique characteristics of natural science courses and social science courses.
Content may be subject to copyright.
RESEARCH ARTICLE
COVID-19’s impacts on the scope,
effectiveness, and interaction characteristics
of online learning: A social network analysis
Junyi Zhang
1‡
, Yigang Ding
1,2‡
, Xinru Yang
1
, Jinping Zhong
1
, XinXin Qiu
1
, Zhishan Zou
3
,
Yujie Xu
1
, Xiunan Jin
1
, Xiaomin Wu
1
, Jingxiu Huang
1
*, Yunxiang ZhengID
1
*
1School of Educational Information Technology, South China Normal University, Guangzhou, Guangdong,
China, 2Hangzhou Zhongce Vocational School Qiantang, Hangzhou, Zhejiang, China, 3Faculty of
Education, Shenzhen University, Shenzhen, Guangdong, China
JZ and YD are contributed equally to this work as first authors.
*jimsow@163.com (JH); dr.zheng.scnu@hotmail.com (YZ)
Abstract
The COVID-19 outbreak brought online learning to the forefront of education. Scholars have
conducted many studies on online learning during the pandemic, but only a few have per-
formed quantitative comparative analyses of students’ online learning behavior before and
after the outbreak. We collected review data from China’s massive open online course plat-
form called icourse.163 and performed social network analysis on 15 courses to explore
courses’ interaction characteristics before, during, and after the COVID-19 pan-demic. Spe-
cifically, we focused on the following aspects: (1) variations in the scale of online learning
amid COVID-19; (2a) the characteristics of online learning interaction during the pandemic;
(2b) the characteristics of online learning interaction after the pandemic; and (3) differences
in the interaction characteristics of social science courses and natural science courses.
Results revealed that only a small number of courses witnessed an uptick in online interac-
tion, suggesting that the pandemic’s role in promoting the scale of courses was not signifi-
cant. During the pandemic, online learning interaction became more frequent among course
network members whose interaction scale increased. After the pandemic, although the
scale of interaction declined, online learning interaction became more effective. The scale
and level of interaction in Electrodynamics (a natural science course) and Economics (a
social science course) both rose during the pan-demic. However, long after the pandemic,
the Economics course sustained online interaction whereas interaction in the Electrodynam-
ics course steadily declined. This discrepancy could be due to the unique characteristics of
natural science courses and social science courses.
1. Introduction
The development of the mobile internet has spurred rapid advances in online learning, offer-
ing novel prospects for teaching and learning and a learning experience completely different
from traditional instruction. Online learning harnesses the advantages of network technology
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 1 / 21
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Zhang J, Ding Y, Yang X, Zhong J, Qiu X,
Zou Z, et al. (2022) COVID-19’s impacts on the
scope, effectiveness, and interaction characteristics
of online learning: A social network analysis. PLoS
ONE 17(8): e0273016. https://doi.org/10.1371/
journal.pone.0273016
Editor: Heng Luo, Central China Normal University,
CHINA
Received: April 20, 2022
Accepted: July 29, 2022
Published: August 23, 2022
Copyright: ©2022 Zhang et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: The data underlying
the results presented in the study were
downloaded from https://www.icourse163.org/ and
are now shared fully on Github (https://github.com/
zjyzhangjunyi/dataset-from-icourse163-for-SNA).
These data have no private information and can be
used for academic research free of charge.
Funding: The author(s) received no specific
funding for this work.
Competing interests: The authors have declared
that no competing interests exist.
and multimedia technology to transcend the boundaries of conventional education [1]. Online
courses have become a popular learning mode owing to their flexibility and openness. During
online learning, teachers and students are in different physical locations but interact in multi-
ple ways (e.g., via online forum discussions and asynchronous group discussions). An analysis
of online learning therefore calls for attention to students’ participation. Alqurashi [2] defined
interaction in online learning as the process of constructing meaningful information and
thought exchanges between more than two people; such interaction typically occurs between
teachers and learners, learners and learners, and the course content and learners.
Massive open online courses (MOOCs), a 21st-century teaching mode, have greatly influ-
enced global education. Data released by China’s Ministry of Education in 2020 show that the
country ranks first globally in the number and scale of higher education MOOCs. The
COVID-19 outbreak has further propelled this learning mode, with universities being urged to
leverage MOOCs and other online resource platforms to respond to government’s “School’s
Out, But Class’s On” policy [3]. Besides MOOCs, to reduce in-person gatherings and curb the
spread of COVID-19, various online learning methods have since become ubiquitous [4].
Though Lederman asserted that the COVID-19 outbreak has positioned online learning tech-
nologies as the best way for teachers and students to obtain satisfactory learning experiences
[5], it remains unclear whether the COVID-19 pandemic has encouraged interaction in online
learning, as interactions between students and others play key roles in academic performance
and largely determine the quality of learning experiences [6]. Similarly, it is also unclear what
impact the COVID-19 pandemic has had on the scale of online learning.
Social constructivism paints learning as a social phenomenon. As such, analyzing the social
structures or patterns that emerge during the learning process can shed light on learning-
based interaction [7]. Social network analysis helps to explain how a social network, rooted in
interactions between learners and their peers, guides individuals’ behavior, emotions, and out-
comes. This analytical approach is especially useful for evaluating interactive relationships
between network members [8]. Mohammed cited social network analysis (SNA) as a method
that can provide timely information about students, learning communities and interactive net-
works. SNA has been applied in numerous fields, including education, to identify the number
and characteristics of interelement relationships. For example, Lee et al. also used SNA to
explore the effects of blogs on peer relationships [7]. Therefore, adopting SNA to examine
interactions in online learning communities during the COVID-19 pandemic can uncover
potential issues with this online learning model.
Taking China’s icourse.163 MOOC platform as an example, we chose 15 courses with a large
number of participants for SNA, focusing on learners’ interaction characteristics before, during,
and after the COVID-19 outbreak. We visually assessed changes in the scale of network interac-
tion before, during, and after the outbreak along with the characteristics of interaction in Gephi.
Examining students’ interactions in different courses revealed distinct interactive network char-
acteristics, the pandemic’s impact on online courses, and relevant suggestions. Findings are
expected to promote effective interaction and deep learning among students in addition to serv-
ing as a reference for the development of other online learning communities.
2. Literature review and research questions
Interaction is deemed as central to the educational experience and is a major focus of research
on online learning. Moore began to study the problem of interaction in distance education as
early as 1989. He defined three core types of interaction: student–teacher, student–content,
and student–student [9]. Lear et al. [10] described an interactivity/ community-process model
of distance education: they specifically discussed the relationships between interactivity,
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 2 / 21
community awareness, and engaging learners and found interactivity and community aware-
ness to be correlated with learner engagement. Zulfikar et al. [11] suggested that discussions
initiated by the students encourage more students’ engagement than discussions initiated by
the instructors. It is most important to afford learners opportunities to interact purposefully
with teachers, and improving the quality of learner interaction is crucial to fostering profound
learning [12]. Interaction is an important way for learners to communicate and share informa-
tion, and a key factor in the quality of online learning [13].
Timely feedback is the main component of online learning interaction. Woo and Reeves
discovered that students often become frustrated when they fail to receive prompt feedback
[14]. Shelley et al. conducted a three-year study of graduate and undergraduate students’ satis-
faction with online learning at universities and found that interaction with educators and stu-
dents is the main factor affecting satisfaction [15]. Teachers therefore need to provide students
with scoring justification, support, and constructive criticism during online learning. Some
researchers examined online learning during the COVID-19 pandemic. They found that most
students preferred face-to-face learning rather than online learning due to obstacles faced
online, such as a lack of motivation, limited teacher-student interaction, and a sense of isola-
tion when learning in different times and spaces [16,17]. However, it can be reduced by
enhancing the online interaction between teachers and students [18].
Research showed that interactions contributed to maintaining students’ motivation to con-
tinue learning [19]. Baber argued that interaction played a key role in students’ academic perfor-
mance and influenced the quality of the online learning experience [20]. Hodges et al.
maintained that well-designed online instruction can lead to unique teaching experiences [21].
Banna et al. mentioned that using discussion boards, chat sessions, blogs, wikis, and other tools
could promote student interaction and improve participation in online courses [22]. During the
COVID-19 pandemic, Mahmood proposed a series of teaching strategies suitable for distance
learning to improve its effectiveness [23]. Lapitan et al. devised an online strategy to ease the tran-
sition from traditional face-to-face instruction to online learning [24]. The preceding discussion
suggests that online learning goes beyond simply providing learning resources; teachers should
ideally design real-life activities to give learners more opportunities to participate.
As mentioned, COVID-19 has driven many scholars to explore the online learning environ-
ment. However, most have ignored the uniqueness of online learning during this time and
have rarely compared pre- and post-pandemic online learning interaction. Taking China’s
icourse.163 MOOC platform as an example, we chose 15 courses with a large number of par-
ticipants for SNA, centering on student interaction before and after the pandemic. Gephi was
used to visually analyze changes in the scale and characteristics of network interaction. The fol-
lowing questions were of particular interest:
(1) Can the COVID-19 pandemic promote the expansion of online learning?
(2a) What are the characteristics of online learning interaction during the pandemic?
(2b) What are the characteristics of online learning interaction after the pandemic?
(3) How do interaction characteristics differ between social science courses and natural science
courses?
3. Methodology
3.1 Research context
We selected several courses with a large number of participants and extensive online interac-
tion among hundreds of courses on the icourse.163 MOOC platform. These courses had been
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 3 / 21
offered on the platform for at least three semesters, covering three periods (i.e., before, during,
and after the COVID-19 outbreak). To eliminate the effects of shifts in irrelevant variables
(e.g., course teaching activities), we chose several courses with similar teaching activities and
compared them on multiple dimensions. All course content was taught online. The teachers of
each course posted discussion threads related to learning topics; students were expected to
reply via comments. Learners could exchange ideas freely in their responses in addition to ask-
ing questions and sharing their learning experiences. Teachers could answer students’ ques-
tions as well. Conversations in the comment area could partly compensate for a relative
absence of online classroom interaction. Teacher–student interaction is conducive to the for-
mation of a social network structure and enabled us to examine teachers’ and students’ learn-
ing behavior through SNA. The comment areas in these courses were intended for learners to
construct knowledge via reciprocal communication. Meanwhile, by answering students’ ques-
tions, teachers could encourage them to reflect on their learning progress. These courses’ suc-
cessive terms also spanned several phases of COVID-19, allowing us to ascertain the
pandemic’s impact on online learning.
3.2 Data collection and preprocessing
To avoid interference from invalid or unclear data, the following criteria were applied to select
representative courses: (1) generality (i.e., public courses and professional courses were chosen
from different schools across China); (2) time validity (i.e., courses were held before during,
and after the pandemic); and (3) notability (i.e., each course had at least 2,000 participants).
We ultimately chose 15 courses across the social sciences and natural sciences (see Table 1).
The coding is used to represent the course name.
To discern courses’ evolution during the pandemic, we gathered data on three terms before,
during, and after the COVID-19 outbreak in addition to obtaining data from two terms com-
pleted well before the pandemic and long after. Our final dataset comprised five sets of interac-
tive data. Finally, we collected about 120,000 comments for SNA. Because each course had a
different start time—in line with fluctuations in the number of confirmed COVID-19 cases in
China and the opening dates of most colleges and universities—we divided our sample into
five phases: well before the pandemic (Phase I); before the pandemic (Phase ); during the
Table 1. List of the target courses.
Course type Course name Coding
Social science courses Educational Wisdom from the Analects of Confucius SS1
Personality Psychology SS2
The World’s Three Major Religions and Arts SS3
Legal Methodology SS4
Economics SS5
Principles and Methods of Instructional Design SS6
Approaching Marx SS7
Natural science courses Electrodynamics NS1
Java Language Programming NS2
College Physics I NS3
Advanced Mathematics I NS4
Machine Learning NS5
Pharmaceutical Chemistry NS6
Medical Statistics NS7
Cell Biology NS8
https://doi.org/10.1371/journal.pone.0273016.t001
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 4 / 21
pandemic (Phase ); after the pandemic (Phase ); and long after the pandemic (Phase ).
We sought to preserve consistent time spans to balance the amount of data in each period
(Fig 1).
3.3 Instrumentation
Participants’ comments and “thumbs-up” behavior data were converted into a network struc-
ture and compared using social network analysis (SNA). Network analysis, according to
M’Chirgui, is an effective tool for clarifying network relationships by employing sophisticated
techniques [25]. Specifically, SNA can help explain the underlying relationships among team
members and provide a better understanding of their internal processes. Yang and Tang used
SNA to discuss the relationship between team structure and team performance [26]. Golbeck
argued that SNA could improve the understanding of students’ learning processes and reveal
learners’ and teachers’ role dynamics [27].
To analyze Question (1), the number of nodes and diameter in the generated network were
deemed as indicators of changes in network size. Social networks are typically represented as
graphs with nodes and degrees, and node count indicates the sample size [15]. Wellman et al.
proposed that the larger the network scale, the greater the number of network members pro-
viding emotional support, goods, services, and companionship [28]. Jan’s study measured the
network size by counting the nodes which represented students, lecturers, and tutors [29].
Similarly, network nodes in the present study indicated how many learners and teachers par-
ticipated in the course, with more nodes indicating more participants. Furthermore, we inves-
tigated the network diameter, a structural feature of social networks, which is a common
Fig 1. Number of confirmed cases in China.
https://doi.org/10.1371/journal.pone.0273016.g001
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 5 / 21
metric for measuring network size in SNA [30]. The network diameter refers to the longest
path between any two nodes in the network. There has been evidence that a larger network
diameter leads to greater spread of behavior [31]. Likewise, Gas
ˇevićet al. found that larger net-
works were more likely to spread innovative ideas about educational technology when analyz-
ing MOOC-related research citations [32]. Therefore, we employed node count and network
diameter to measure the network’s spatial size and further explore the expansion characteristic
of online courses. Brief introduction of these indicators can be summarized in Table 2.
To address Question (2), a list of interactive analysis metrics in SNA were introduced to
scrutinize learners’ interaction characteristics in online learning during and after the pan-
demic, as shown below:
1. The average degree reflects the density of the network by calculating the average number of
connections for each node. As Rong and Xu suggested, the average degree of a network
indicates how active its participants are [33]. According to Hu, a higher average degree
implies that more students are interacting directly with each other in a learning context
[34]. The present study inherited the concept of the average degree from these previous
studies: the higher the average degree, the more frequent the interaction between individu-
als in the network.
2. Essentially, a weighted average degree in a network is calculated by multiplying each degree
by its respective weight, and then taking the average. Bydz
ˇovska
´took the strength of the
relationship into account when determining the weighted average degree [35]. By calculat-
ing friendship’s weighted value, Maroulis assessed peer achievement within a small-school
reform [36]. Accordingly, we considered the number of interactions as the weight of the
degree, with a higher average degree indicating more active interaction among learners.
3. Network density is the ratio between actual connections and potential connections in a net-
work. The more connections group members have with each other, the higher the network
density. In SNA, network density is similar to group cohesion, i.e., a network of more
strong relationships is more cohesive [37]. Network density also reflects how much all
members are connected together [38]. Therefore, we adopted network density to indicate
the closeness among network members. Higher network density indicates more frequent
interaction and closer communication among students.
4. Clustering coefficient describes local network attributes and indicates that two nodes in the
network could be connected through adjacent nodes. The clustering coefficient measures
users’ tendency to gather (cluster) with others in the network: the higher the clustering coef-
ficient, the more frequently users communicate with other group members. We regarded
this indicator as a reflection of the cohesiveness of the group [39].
5. In a network, the average path length is the average number of steps along the shortest
paths between any two nodes. Oliveres has observed that when an average path length is
small, the route from one node to another is shorter when graphed [40]. This is especially
true in educational settings where students tend to become closer friends. So we consider
Table 2. Introduction of indicators for Question (1).
Indicator Definition Study
Node count The number of learners and teachers who have participated in
the course.
Cole et al. [15]; Jan [29]
Network
diameter
The longest path between any two nodes in the network. Serrat O [30]; Gas
ˇevićet al.
[32]
https://doi.org/10.1371/journal.pone.0273016.t002
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 6 / 21
that the smaller the average path length, the greater the possibility of interaction between
individuals in the network.
6. A network with a large number of nodes, but whose average path length is surprisingly
small, is known as the small-world effect [41]. A higher clustering coefficient and shorter
average path length are important indicators of a small-world network: a shorter average
path length enables the network to spread information faster and more accurately; a higher
clustering coefficient can promote frequent knowledge exchange within the group while
boosting the timeliness and accuracy of knowledge dissemination [42]. Brief introduction
of these indicators can be summarized in Table 3.
To analyze Question 3, we used the concept of closeness centrality, which determines how close
a vertex is to others in the network. As Opsahl et al. explained, closeness centrality reveals how
closely actors are coupled with their entire social network [43]. In order to analyze social network-
based engineering education, Putnik et al. examined closeness centrality and found that it was sig-
nificantly correlated with grades [38]. We used closeness centrality to measure the position of an
individual in the network. Brief introduction of these indicators can be summarized in Table 4.
3.4 Ethics statement
This study was approved by the Academic Committee Office (ACO) of South China Normal
University (http://fzghb.scnu.edu.cn/), Guangzhou, China. Research data were collected from
the open platform and analyzed anonymously. There are thus no privacy issues involved in
this study.
4. Results
4.1 COVID-19’s role in promoting the scale of online courses was not as
important as expected
As shown in Fig 2, the number of course participants and nodes are closely correlated with the
pandemic’s trajectory. Because the number of participants in each course varied widely, we
Table 3. Introduction of indicators for Question (2).
Indicators Definition Study
Average degree The density of the network by calculating the average number
of connections for each node.
Rong & Xu [33]; Hu [34]
Weighted average
degree
Calculating the average degree based on the network’s
different weights.
Bydz
ˇovska
´[35]; Maroulis
[36]
Network density The ratio between actual connections and potential
connections in a network.
Wise S [37]; Potts BB [38]
Clustering
coefficient
Two nodes in the network could be connected through
adjacent nodes.
Grunspan DZ et al. [39]
Average path
length
The average number of steps along the shortest paths between
any two nodes.
Oliveres [40]
Small-world effect A network with a large number of nodes, but whose average
path length is surprisingly small.
Travers J et al. [41]; Watts
DJ et al. [42]
https://doi.org/10.1371/journal.pone.0273016.t003
Table 4. Introduction of indicators for Question (3).
Indicators Definition Study
closeness centrality Measure the position of an individual in the network. Opsahl [44];
Putnik [45]
https://doi.org/10.1371/journal.pone.0273016.t004
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 7 / 21
normalized the number of participants and nodes to more conveniently visualize course
trends. Fig 2 depicts changes in the chosen courses’ number of participants and nodes before
the pandemic (Phase II), during the pandemic (Phase III), and after the pandemic (Phase IV).
The number of participants in most courses during the pandemic exceeded those before and
after the pandemic. But the number of people who participate in interaction in some courses
did not increase.
In order to better analyze the trend of interaction scale in online courses before, during,
and after the pandemic, the selected courses were categorized according to their scale change.
When the number of participants increased (decreased) beyond 20% (statistical experience)
and the diameter also increased (decreased), the course scale was determined to have increased
Fig 2. Participants and nodes in course network.
https://doi.org/10.1371/journal.pone.0273016.g002
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 8 / 21
(decreased); otherwise, no significant change was identified in the course’s interaction scale.
Courses were subsequently divided into three categories: increased interaction scale, decreased
interaction scale, and no significant change. Results appear in Table 5.
From before the pandemic until it broke out, the interaction scale of five courses increased,
accounting for 33.3% of the full sample; one course’s interaction scale declined, accounting for
6.7%. The interaction scale of nine courses decreased, accounting for 60%. The pandemic’s
role in promoting online courses thus was not as important as anticipated, and most courses’
interaction scale did not change significantly throughout.
No courses displayed growing interaction scale after the pandemic: the interaction scale of
nine courses fell, accounting for 60%; and the interaction scale of six courses did not shift sig-
nificantly, accounting for 40%. Courses with an increased scale of interaction during the pan-
demic did not maintain an upward trend. On the contrary, the improvement in the pandemic
caused learners’ enthusiasm for online learning to wane. We next analyzed several interaction
metrics to further explore course interaction during different pandemic periods.
4.2 Characteristics of online learning interaction amid COVID-19
4.2.1 During the COVID-19 pandemic, online learning interaction in some courses
became more active. Changes in course indicators with the growing interaction scale during
the pandemic are presented in Fig 3, including SS5, SS6, NS1, NS3, and NS8. The horizontal
ordinate indicates the number of courses, with red color representing the rise of the indicator
value on the vertical ordinate and blue representing the decline.
Specifically: (1) The average degree and weighted average degree of the five course networks
demonstrated an upward trend. The emergence of the pandemic promoted students’ enthusi-
asm; learners were more active in the interactive network. (2) Fig 3 shows that 3 courses had
increased network density and 2 courses had decreased. The higher the network density, the
more communication within the team. Even though the pandemic accelerated the interaction
scale and frequency, the tightness between learners in some courses did not improve. (3) The
clustering coefficient of social science courses rose whereas the clustering coefficient and
small-world property of natural science courses fell. The higher the clustering coefficient and
the small-world property, the better the relationship between adjacent nodes and the higher
the cohesion [39]. (4) Most courses’ average path length increased as the interaction scale
increased. However, when the average path length grew, adverse effects could manifest: com-
munication between learners might be limited to a small group without multi-directional
interaction.
When the pandemic emerged, the only declining network scale belonged to a natural sci-
ence course (NS2). The change in each course index is pictured in Fig 4. The abscissa indicates
the size of the value, with larger values to the right. The red dot indicates the index value before
the pandemic; the blue dot indicates its value during the pandemic. If the blue dot is to the
right of the red dot, then the value of the index increased; otherwise, the index value declined.
Table 5. Courses’ scale change.
Period Course scale Course coding (Social sciences) Course coding (Natural sciences)
Before COVID-19 pandemic–During COVID-19 pandemic Increase SS5; SS6 NS1; NS3; NS8
Decrease - NS2
No significant change SS1; SS2; SS3; SS4; SS7 NS4; NS5; NS6; NS7
During COVID-19 pandemic–After COVID-19 pandemic Decrease SS1; SS2; SS3; SS6; SS7 NS2; NS3; NS7; NS8
No significant change SS4; SS5 NS1; NS4; NS5; NS6
https://doi.org/10.1371/journal.pone.0273016.t005
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 9 / 21
Only the weighted average degree of the course network increased. The average degree, net-
work density decreased, indicating that network members were not active and that learners’
interaction degree and communication frequency lessened. Despite reduced learner interac-
tion, the average path length was small and the connectivity between learners was adequate.
Fig 3. Indicator changes in courses with increased interaction scale from beforethe COVID-19 outbreak until the outbreak.
https://doi.org/10.1371/journal.pone.0273016.g003
Fig 4. Indicator changes in courses with decreased interaction scale from beforethe COVID-19 pandemic until the outbreak.
https://doi.org/10.1371/journal.pone.0273016.g004
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 10 / 21
4.2.2 After the COVID-19 pandemic, the scale decreased rapidly, but most course inter-
action was more effective. Fig 5 shows the changes in various courses’ interaction indicators
after the pandemic, including SS1, SS2, SS3, SS6, SS7, NS2, NS3, NS7, and NS8.
Specifically: (1) The average degree and weighted average degree of most course networks
decreased. The scope and intensity of interaction among network members declined rapidly,
as did learners’ enthusiasm for communication. (2) The network density of seven courses also
fell, indicating weaker connections between learners in most courses. (3) In addition, the clus-
tering coefficient and small-world property of most course networks decreased, suggesting lit-
tle possibility of small groups in the network. The scope of interaction between learners was
not limited to a specific space, and the interaction objects had no significant tendencies. (4)
Although the scale of course interaction became smaller in this phase, the average path length
of members’ social networks shortened in nine courses. Its shorter average path length would
expedite the spread of information within the network as well as communication and sharing
among network members.
Fig 6 displays the evolution of course interaction indicators without significant changes in
interaction scale after the pandemic, including SS4, SS5, NS1, NS4, NS5, and NS6.
Specifically: (1) Some course members’ social networks exhibited an increase in the average
and weighted average. In these cases, even though the course network’s scale did not continue
to increase, communication among network members rose and interaction became more fre-
quent and deeper than before. (2) Network density and average path length are indicators of
social network density. The greater the network density, the denser the social network; the
shorter the average path length, the more concentrated the communication among network
Fig 5. Changes in indicators of courses with decreased interaction scale from the COVID-19 outbreak to after the pandemic.
https://doi.org/10.1371/journal.pone.0273016.g005
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 11 / 21
members. However, at this phase, the average path length and network density in most courses
had increased. Yet the network density remained small despite having risen (Table 6). Even
with more frequent learner interaction, connections remained distant and the social network
was comparatively sparse.
In summary, the scale of interaction did not change significantly overall. Nonetheless, some
course members’ frequency and extent of interaction increased, and the relationships between
network members became closer as well. In the study, we found it interesting that the interac-
tion scale of Economics (a social science course) course and Electrodynamics (a natural sci-
ence course) course expanded rapidly during the pandemic and retained their interaction
scale thereafter. We next assessed these two courses to determine whether their level of interac-
tion persisted after the pandemic.
4.3 Analyses of natural science courses and social science courses
4.3.1 Analyses of the interaction characteristics of Economics and Electrodynamics.
Economics and Electrodynamics are social science courses and natural science courses, respec-
tively. Members’ interaction within these courses was similar: the interaction scale increased
significantly when COVID-19 broke out (Phase ), and no significant changes emerged after
the pandemic (Phase ). We hence focused on course interaction long after the outbreak
(Phase V) and compared changes across multiple indicators, as listed in Table 7.
As the pandemic continued to improve, the number of participants and the diameter long
after the outbreak (Phase V) each declined for Economics compared with after the pandemic
(Phase IV). The interaction scale decreased, but the interaction between learners was much
Fig 6. Changes in indicators of courses with no change in interaction scale from the COVID-19outbreak until after the pandemic.
https://doi.org/10.1371/journal.pone.0273016.g006
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 12 / 21
deeper. Specifically: (1) The weighted average degree, network density, clustering coefficient, and
small-world property each reflected upward trends. The pandemic therefore exerted a strong
impact on this course. Interaction was well maintained even after the pandemic. The smaller net-
work scale promoted members’ interaction and communication. (2) Compared with after the
pandemic (Phase IV), members’ network density increased significantly, showing that relation-
ships between learners were closer and that cohesion was improving. (3) At the same time, as the
clustering coefficient and small-world property grew, network members demonstrated strong
small-group characteristics: the communication between them was deepening and their enthusi-
asm for interaction was higher. (4) Long after the COVID-19 outbreak (Phase V), the average
path length was reduced compared with previous terms, knowledge flowed more quickly among
network members, and the degree of interaction gradually deepened.
The average degree, weighted average degree, network density, clustering coefficient, and
small-world property of Electrodynamics all decreased long after the COVID-19 outbreak
(Phase V) and were lower than during the outbreak (Phase ). The level of learner interaction
therefore gradually declined long after the outbreak (Phase V), and connections between
learners were no longer active. Although the pandemic increased course members’ extent of
interaction, this rise was merely temporary: students’ enthusiasm for learning waned rapidly
and their interaction decreased after the pandemic (Phase IV). To further analyze the interac-
tion characteristics of course members in Economics and Electrodynamics, we evaluated the
closeness centrality of their social networks, as shown in section 4.3.2.
4.3.2 Analysis of the closeness centrality of Economics and Electrodynamics. The
change in the closeness centrality of social networks in Economics was small, and no sharp
upward trend appeared during the pandemic outbreak, as shown in Fig 7. The emergence of
COVID-19 apparently fostered learners’ interaction in Economics albeit without a significant
impact. The closeness centrality changed in Electrodynamics varied from that of Economics:
upon the COVID-19 outbreak, closeness centrality was significantly different from other
semesters. Communication between learners was closer and interaction was more effective.
Electrodynamics course members’ social network proximity decreased rapidly after the pan-
demic. Learners’ communication lessened. In general, Economics course showed better inter-
action before the outbreak and was less affected by the pandemic; Electrodynamics course was
Table 6. Network density.
Phase SS4 SS5 NS1 NS4 NS5 NS6
Phase 0.05 0.033 0.043 0.003 0.048 0.024
Phase 0.077 0.031 0.038 0.002 0.04 0.043
Phase 0.5 0.033 0.004 0.003 0.039 0.013
Phase 0.107 0.038 0.098 0.004 0.111 0.041
Phase 0.083 0.083 0.058
https://doi.org/10.1371/journal.pone.0273016.t006
Table 7. Comparison of various indicators for Economics and Electrodynamics.
Course Phase Number of
participants
Diameter Average
degree
Weighted
average
Network
density
Clustering
coefficient
Small-world
property
Average path
length
Economics (SS5) 12382 4 0.893 1.357 0.033 0.057 0.035 1.615
8891 6 1.1 2.133 0.038 0.017 0.007 2.321
6160 1 1 2.308 0.083 0.104 0.104 1
Electrodynamics
(NS1)
4685 5 1.142 5.301 0.004 0.041 0.023 1.753
3722 2 1.667 4.444 0.098 0.2 0.111 1.805
2170 3 1 2.462 0.083 0.01 0.006 1.812
https://doi.org/10.1371/journal.pone.0273016.t007
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 13 / 21
more affected by the pandemic and showed different interaction characteristics at different
periods of the pandemic.
5. Discussion
We referred to discussion forums from several courses on the icourse.163 MOOC platform to
compare online learning before, during, and after the COVID-19 pandemic via SNA and to
Fig 7. Closeness centrality of Economics and Electrodynamics. (Note: "����" indicates the significant distinction in closeness centrality between the two
periods, otherwise no significant distinction).
https://doi.org/10.1371/journal.pone.0273016.g007
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 14 / 21
delineate the pandemic’s effects on online courses. Only 33.3% of courses in our sample
increased in terms of interaction during the pandemic; the scale of interaction did not rise in
any courses thereafter. When the courses scale rose, the scope and frequency of interaction
showed upward trends during the pandemic; and the clustering coefficient of natural science
courses and social science courses differed: the coefficient for social science courses tended to
rise whereas that for natural science courses generally declined. When the pandemic broke
out, the interaction scale of a single natural science course decreased along with its interaction
scope and frequency. The amount of interaction in most courses shrank rapidly during the
pandemic and network members were not as active as they had been before. However, after
the pandemic, some courses saw declining interaction but greater communication between
members; interaction also became more frequent and deeper than before.
5.1 During the COVID-19 pandemic, the scale of interaction increased in
only a few courses
The pandemic outbreak led to a rapid increase in the number of participants in most courses;
however, the change in network scale was not significant. The scale of online interaction
expanded swiftly in only a few courses; in others, the scale either did not change significantly or
displayed a downward trend. After the pandemic, the interaction scale in most courses decreased
quickly; the same pattern applied to communication between network members. Learners’ enthu-
siasm for online interaction reduced as the circumstances of the pandemic improved—potentially
because, during the pandemic, China’s Ministry of Education declared “School’s Out, But Class’s
On” policy. Major colleges and universities were encouraged to use the Internet and informa-
tional resources to provide learning support, hence the sudden increase in the number of partici-
pants and interaction in online courses [46]. After the pandemic, students’ enthusiasm for online
learning gradually weakened, presumably due to easing of the pandemic [47]. More activities also
transitioned from online to offline, which tempered learners’ online discussion. Research has
shown that long-term online learning can even bore students [48].
Most courses’ interaction scale decreased significantly after the pandemic. First, teachers
and students occupied separate spaces during the outbreak, had few opportunities for mutual
cooperation and friendship, and lacked a sense of belonging [49]. Students’ enthusiasm for
learning dissipated over time [50]. Second, some teachers were especially concerned about
adapting in-person instructional materials for digital platforms; their pedagogical methods
were ineffective, and they did not provide learning activities germane to student interaction
[51]. Third, although teachers and students in remote areas were actively engaged in online
learning, some students could not continue to participate in distance learning due to inade-
quate technology later in the outbreak [52].
5.2 Characteristics of online learning interaction during and after the
COVID-19 pandemic
5.2.1 During the COVID-19 pandemic, online interaction in most courses did not
change significantly. The interaction scale of only a few courses increased during the pan-
demic. The interaction scope and frequency of these courses climbed as well. Yet even as the
degree of network interaction rose, course network density did not expand in all cases. The
pandemic sparked a surge in the number of online learners and a rapid increase in network
scale, but students found it difficult to interact with all learners. Yau pointed out that a greater
network scale did not enrich the range of interaction between individuals; rather, the number
of individuals who could interact directly was limited [53]. The internet facilitates
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 15 / 21
interpersonal communication. However, not everyone has the time or ability to establish close
ties with others [54].
In addition, social science courses and natural science courses in our sample revealed dispa-
rate trends in this regard: the clustering coefficient of social science courses increased and that
of natural science courses decreased. Social science courses usually employ learning
approaches distinct from those in natural science courses [55]. Social science courses empha-
size critical and innovative thinking along with personal expression [56]. Natural science
courses focus on practical skills, methods, and principles [57]. Therefore, the content of social
science courses can spur large-scale discussion among learners. Some course evaluations indi-
cated that the course content design was suboptimal as well: teachers paid close attention to
knowledge transmission and much less to piquing students’ interest in learning. In addition,
the thread topics that teachers posted were scarcely diversified and teachers’ questions lacked
openness. These attributes could not spark active discussion among learners.
5.2.2 Online learning interaction declined after the COVID-19 pandemic. Most
courses’ interaction scale and intensity decreased rapidly after the pandemic, but some did not
change. Courses with a larger network scale did not continue to expand after the outbreak,
and students’ enthusiasm for learning paled. The pandemic’s reduced severity also influenced
the number of participants in online courses. Meanwhile, restored school order moved many
learning activities from virtual to in-person spaces. Face-to-face learning has gradually
replaced online learning, resulting in lower enrollment and less interaction in online courses.
Prolonged online courses could have also led students to feel lonely and to lack a sense of
belonging [58].
The scale of interaction in some courses did not change substantially after the pandemic yet
learners’ connections became tighter. We hence recommend that teachers seize pandemic-
related opportunities to design suitable activities. Additionally, instructors should promote
student-teacher and student-student interaction, encourage students to actively participate
online, and generally intensify the impact of online learning.
5.3 What are the characteristics of interaction in social science courses and
natural science courses?
The level of interaction in Economics (a social science course) was significantly higher than
that in Electrodynamics (a natural science course), and the small-world property in Economics
increased as well. To boost online courses’ learning-related impacts, teachers can divide groups
of learners based on the clustering coefficient and the average path length. Small groups of stu-
dents may benefit teachers in several ways: to participate actively in activities intended to
expand students’ knowledge, and to serve as key actors in these small groups. Cultivating stu-
dents’ keenness to participate in class activities and self-management can also help teachers
guide learner interaction and foster deep knowledge construction.
As evidenced by comments posted in the Electrodynamics course, we observed less interac-
tion between students. Teachers also rarely urged students to contribute to conversations.
These trends may have arisen because teachers and students were in different spaces. Teachers
might have struggled to discern students’ interaction status. Teachers could also have failed to
intervene in time, to design online learning activities that piqued learners’ interest, and to
employ sound interactive theme planning and guidance. Teachers are often active in tradi-
tional classroom settings. Their roles are comparatively weakened online, such that they pos-
sess less control over instruction [59]. Online instruction also requires a stronger hand in
learning: teachers should play a leading role in regulating network members’ interactive com-
munication [60]. Teachers can guide learners to participate, help learners establish social
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 16 / 21
networks, and heighten students’ interest in learning [61]. Teachers should attend to core
members in online learning while also considering edge members; by doing so, all network
members can be driven to share their knowledge and become more engaged. Finally, teachers
and assistant teachers should help learners develop knowledge, exchange topic-related ideas,
pose relevant questions during course discussions, and craft activities that enable learners to
interact online [62]. These tactics can improve the effectiveness of online learning.
As described, network members displayed distinct interaction behavior in Economics and
Electrodynamics courses. First, these courses varied in their difficulty: the social science course
seemed easier to understand and focused on divergent thinking. Learners were often willing to
express their views in comments and to ponder others’ perspectives [63]. The natural science
course seemed more demanding and was oriented around logical thinking and skills [64]. Sec-
ond, courses’ content differed. In general, social science courses favor the acquisition of declar-
ative knowledge and creative knowledge compared with natural science courses. Social science
courses also entertain open questions [65]. Natural science courses revolve around principle
knowledge, strategic knowledge, and transfer knowledge [66]. Problems in these courses are
normally more complicated than those in social science courses. Third, the indicators affecting
students’ attitudes toward learning were unique. Guo et al. discovered that “teacher feedback”
most strongly influenced students’ attitudes towards learning social science courses but had
less impact on students in natural science courses [67]. Therefore, learners in social science
courses likely expect more feedback from teachers and greater interaction with others.
6. Conclusion and future work
Our findings show that the network interaction scale of some online courses expanded during
the COVID-19 pandemic. The network scale of most courses did not change significantly,
demonstrating that the pandemic did not notably alter the scale of course interaction. Online
learning interaction among course network members whose interaction scale increased also
became more frequent during the pandemic. Once the outbreak was under control, although
the scale of interaction declined, the level and scope of some courses’ interactive networks con-
tinued to rise; interaction was thus particularly effective in these cases. Overall, the pandemic
appeared to have a relatively positive impact on online learning interaction. We considered a
pair of courses in detail and found that Economics (a social science course) fared much better
than Electrodynamics (a natural science course) in classroom interaction; learners were more
willing to partake in-class activities, perhaps due to these courses’ unique characteristics. Brint
et al. also came to similar conclusions [57].
This study was intended to be rigorous. Even so, several constraints can be addressed in
future work. The first limitation involves our sample: we focused on a select set of courses
hosted on China’s icourse.163 MOOC platform. Future studies should involve an expansive
collection of courses to provide a more holistic understanding of how the pandemic has influ-
enced online interaction. Second, we only explored the interactive relationship between learn-
ers and did not analyze interactive content. More in-depth content analysis should be carried
out in subsequent research. All in all, the emergence of COVID-19 has provided a new path
for online learning and has reshaped the distance learning landscape. To cope with associated
challenges, educational practitioners will need to continue innovating in online instructional
design, strengthen related pedagogy, optimize online learning conditions, and bolster teachers’
and students’ competence in online learning.
Author Contributions
Conceptualization: Jingxiu Huang, Yunxiang Zheng.
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 17 / 21
Data curation: Junyi Zhang, Yigang Ding, Xinru Yang, Jinping Zhong, XinXin Qiu, Xiunan
Jin, Xiaomin Wu.
Formal analysis: Junyi Zhang, Yigang Ding.
Methodology: Junyi Zhang, Yigang Ding.
Supervision: Jingxiu Huang, Yunxiang Zheng.
Writing original draft: Yigang Ding, Xinru Yang, Zhishan Zou, Yujie Xu.
Writing review & editing: Junyi Zhang, Jingxiu Huang, Yunxiang Zheng.
References
1. Lin MH, Chen HC, Liu KS. A Study of the Effects of Digital Learning on Learning Motivation and Learn-
ing Outcome. Eurasia Journal of Mathematics Science and Technology Education. 2017; 13(7):3553–
3564. https://doi.org/10.12973/eurasia.2017.00744a
2. Alqurashi E. Predicting student satisfaction and perceived learning within online learning environments.
Distance Education. 2019 Jan 2; 40(1):133–148. https://doi.org/10.1080/01587919.2018.1553562
3. Zhou L, Li F., Wu S., & Zhou M. “School’s out, but class’s on”, the largest online education in the world
today: taking China’s practical exploration during the COVID-19 epidemic prevention and control as an
example. Best Evidence in Chinese Education. 2020; 4(2):501–519. https://doi.org/10.15354/bece.20.
ar023
4. Wang C, Zhao H. The Impact of COVID-19 on Anxiety in Chinese University Students. Frontiers in Psy-
chology. 2020; 11. https://doi.org/10.3389/fpsyg.2020.01168 PMID: 32574244
5. Lederman D. Will shift to remote teaching be boon or bane for online learning. Inside Higher Education.
2020; 18.Available from: https://www.insidehighered.com/digital-learning/article/2020/03/18/most-
teaching-going-remote-will-help-or-hurt-online-learning
6. Doran, Rice P, Chad, et al. Social network analysis as a method for analyzing interaction in collaborative
online learning environments. Journal of Systemics, Cybernetics Informatics. 2011; 9.
7. Lee J, Bonk CJ. Social network analysis of peer relationships and online interactions in a blended class
using blogs. The Internet Higher Education. 2016; 28(1):35–44. https://doi.org/10.1016/j.iheduc.2015.
09.001
8. Scott J. Social network analysis. Vol. 22. 2013. (Sociology; 1). https://doi.org/10.1177/
0038038588022001007
9. Moore MG. Three Types of Interaction. Distance Education New Perspectives. 1993; 3(2):1–7. https://
doi.org/10.1080/08923648909526659
10. Lear JL, Ansorge C, Steckelberg A. Interactivity/community process model for the online education
environment. Journal of online learning teaching. 2010; 6(1):71–77.
11. Zulfikar AF, Muhidin A., Suparta W., Trisetyarso A., Abbas B.S. and Kang C.H. The effectiveness of
online learning with facilitation method. Procedia Computer Science. 2019; 161:32–40.
12. Garrison DR, Cleveland-Innes M. Facilitating cognitive presence in online learning: Interaction is not
enough. The American journal of distance education. 2005; 19(3):133–148. https://doi.org/10.1207/
s15389286ajde1903_2
13. Wu B. Influence of MOOC learners discussion forum social interactions on online reviews of MOOC.
Education and Information Technologies. 2021; 26:3483–3496. https://doi.org/10.1007/s10639-020-
10412-z
14. Woo Y, Reeves TC. Meaningful interaction in web-based learning: A social constructivist interpretation.
Internet and Higher Education. 2007 2007; 10(1):15–25. https://doi.org/10.1016/j.iheduc.2006.10.005
15. Cole MT, Shelley DJ, Swartz LB. Online Instruction, E-Learning, and Student Satisfaction: A Three
Year Study. International Review of Research in Open and Distributed Learning. 2014 Dec; 15(6):111–
131. https://doi.org/10.19173/irrodl.v15i6.1748
16. Alawamleh M, Al-Twait LM, Al-Saht GR. The effect of online learning on communication between
instructors and students during Covid-19 pandemic. Asian Education and Development Studies. 2020;
11(2):380–400. https://doi.org/10.1108/AEDS-06-2020-0131
17. Amir LR, Tanti I, Maharani DA, et al. Student perspective of classroom and distance learning during
COVID-19 pandemic in the undergraduate dental study program Universitas Indonesia. Bmc Medical
Education. 2020 Oct 29; 20(1). https://doi.org/10.1186/s12909-020-02312-0 PMID: 33121488
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 18 / 21
18. Chiu TKF, & Hew T. K. F. Factors influencing peer learning and performance in MOOC asynchronous
online discussion forum. Australasian Journal of Educational Technology. 2018; 34(4):16–28. https://
doi.org/10.14742/ajet.3240
19. Zhi Liu NZ, Xian Peng, Sannyuya Liu, Zongkai Yang, Jiyi Peng, Zhu Su, et al. Exploring the Relationship
Between Social Interaction, Cognitive Processing and Learning Achievements in a MOOC Discussion
Forum. Journal of Educational Computing Research. 2021. https://doi.org/10.1177/
07356331211027300
20. Baber H. Determinants of students’ perceived learning outcome and satisfaction in online learning dur-
ing the pandemic of COVID-19. Journal of Education and E-Learning Research. 2020; 7(3):285–292.
21. Hodges CB, Moore S, Lockee BB, et al. The difference between emergency remote teaching and online
learning. Educause. 2020.
22. Banna J, Grace Lin M-F, Stewart M, et al. Interaction matters: Strategies to promote engaged learning
in an online introductory nutrition course. Journal of online learning and teaching. 2015 2015-Jun; 11
(2):249–261.Available from: http://MEDLINE:27441032 PMID: 27441032
23. Mahmood S. Instructional Strategies for Online Teaching in COVID-19 Pandemic. Human Behavior
and Emerging Technologies. 2021 2021; 3(1):199–203. https://doi.org/10.1002/hbe2.218
24. Lapitan LDS Jr., Tiangco CE, Sumalinog DAG, et al. An effective blended online teaching and learning
strategy during the COVID-19 pandemic. Education for Chemical Engineers. 2020 Apr; 35:116–131.
https://doi.org/10.1016/j.ece.2021.01.012
25. M’Chirgui Z. The economics of the smart card industry: towards coopetitive strategies. Economics of
Innovation New Technology. 2005; 14(6):455–477. https://doi.org/10.1080/1043859042000304070
26. Yang HL, Tang HH. Team structure and team performance in IS development: a social network per-
spective. Information and Management. 2004 Jan; 41(3):335–349. https://doi.org/10.1016/s0378-7206
(03)00078-8
27. Golbeck J. Analyzing the Social Web. Chapter 3—Network Structure and Measures 2013.
28. Wellman B, Berkowitz SD. Social Structures: A Network Approach. Vol. 83. 1988. (American Politica-
lence Association; 4). https://doi.org/10.2307/1961708
29. Jan SK, Vlachopoulos P. Social Network Analysis: A Framework for Identifying Communities in Higher
Education Online Learning. Technology, Knowledge and Learning. 2019 Dec.; 24(4):621–39. https://
doi.org/10.1007/s10758-018-9375-y
30. Serrat O. Social network analysis. Knowledge solutions: Springer; 2017. p. 39–43. https://doi.org/10.
1007/978-981-10-0983-9_9
31. Centola D. The Spread of Behavior in an Online Social Network Experiment. Science. 2010; 329
(5996):1194–1197. https://doi.org/10.1126/science.1185231 PMID: 20813952
32. Gas
ˇevićD, KovanovićV, JoksimovićS, et al. Where is research on massive open online courses
headed? A data analysis of the MOOC Research Initiative. International Review of Research in Open
Distributed Learning. 2014; 15(5):134–176. https://doi.org/10.19173/irrodl.v15i5.1954
33. Rong Y, Xu E, editors. Strategies for the Management of the Government Affairs Microblogs in China
Based on the SNA of Fifty Government Affairs Microblogs in Beijing. 14th International Conference on
Service Systems and Service Management 2017.
34. Hu X, Chu S, editors. A comparison on using social media in a professional experience course. Interna-
tional Conference on Social Media and Society; 2013.
35. Bydz
ˇovska
´H. A Comparative Analysis of Techniques for Predicting Student Performance. Proceedings
of the 9th International Conference on Educational Data Mining; Raleigh, NC, USA: International Edu-
cational Data Mining Society2016. p. 306–311.
36. Maroulis S, Gomez LM. Does “connectedness” matter? Evidence from a social network analysis within
a small-school reform. Teachers College Record. 2008; 110(9):1901–1929.Available from: https://
www.tcrecord.org
37. Wise S. Can a team have too much cohesion? The dark side to network density. European manage-
ment journal. 2014; 32(5):703–711.
38. Potts BB. Social network analysis—methods and applications. Vol. 37. 1994. English. (Acta Sociolo-
gica; 4).
39. Grunspan DZ, Wiggins BL, Goodreau SM. Understanding classrooms through social network analysis:
A primer for social network analysis in education research. CBE—Life Sciences Education. 2014; 13
(2):167–178. https://doi.org/10.1187/cbe.13-08-0162 PMID: 26086650
40. Olivares D, Adesope O, Hundhausen C, et al., editors. Using social network analysis to measure the
effect of learning analytics in computing education. 19th IEEE International Conference on Advanced
Learning Technologies 2019.
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 19 / 21
41. Travers J, Milgram S. An experimental study of the small world problem. Social Networks: Elsevier;
1977. p. 179–197. https://doi.org/10.1016/B978-0-12-442450-0.50018–3
42. Watts DJ, Strogatz SH. Collective dynamics of ’small-world’ networks. Nature. 1998; 393(6684):440–
442. https://doi.org/10.1038/30918 PMID: 9623998
43. Okamoto K, Chen W, Li X-Y, editors. Ranking of closeness centrality for large-scale social networks.
International workshop on frontiers in algorithmics; 2008; Springer, Berlin, Heidelberg: Springer.
44. Opsahl T, Agneessens F, Skvoretz J. Node centrality in weighted networks: Generalizing degree and
shortest paths [Article]. Social Networks. 2010 Jul; 32(3):245–251. https://doi.org/10.1016/j.socnet.
2010.03.006 Available from: <Go to ISI>://WOS:000279409000009
45. Putnik G, Costa E, Alves C, et al. Analysing the correlation between social network analysis measures
and performance of students in social network-based engineering education [Article]. International Jour-
nal of Technology and Design Education. 2016 Aug; 26(3):413–437. https://doi.org/10.1007/s10798-
015-9318-z
46. Bao W. COVID-19 and online teaching in higher education: a case study of Peking University. Human
Behavior and Emerging Technologies. 2020. https://doi.org/10.1002/hbe2.191 PMID: 32510042
47. Ding Y, Yang X, Zheng Y, editors. COVID-19’s Effects on the Scope, Effectiveness, and Roles of
Teachers in Online Learning Based on Social Network Analysis: A Case Study. International Confer-
ence on Blended Learning; 2021: Springer.
48. Patricia Aguilera-Hermida A. College students’ use and acceptance of emergency online learning due
to COVID-19. International journal of educational research open. 2020; 1:100011. https://doi.org/10.
1016/j.ijedro.2020.100011 PMID: 35059662
49. De Gagne JC, Walters K. Online teaching experience: A qualitative metasynthesis. MERLOT Journal of
Online Learning Teaching. 2009; 5(4):577.
50. Rajab MH, Gazal AM, Alkattan K. Challenges to Online Medical Education During the COVID-19 Pan-
demic. Cureus. 2020; 12(7):e8966. https://doi.org/10.7759/cureus.8966 PMID: 32766008
51. Aliyyah RR, Rachmadtullah R, Samsudin A, et al. The perceptions of primary school teachers of online
learning during the COVID-19 pandemic period: A case study in Indonesia. Journal of Ethnic Cultural
Studies. 2020; 7(2):90–109. https://doi.org/10.29333/ejecs/388
52. Zhong R. The coronavirus exposes education’s digital divide. The New York Times. 2020. Available
from: https://www.nytimes.com/2020/03/17/technology/china-schools-coronavirus.html
53. Yau JC, Reich SM, Wang Y, et al. More friends, more interactions? The association between network
size and interactions on Facebook. First Monday. 2018; 23(5). https://doi.org/10.5210/fm.v22i5.8195
54. Fung YYH. Collaborative online learning: interaction patterns and limiting factors. Open Learning: The
Journal of Open, Distance and e-Learning. 2004; 19(2):135–149. https://doi.org/10.1080/
0268051042000224743
55. Neumann R. Disciplinary differences and university teaching. Studies in higher education. 2001; 26
(2):135–146. https://doi.org/10.1080/03075070120052071
56. Sarah Schneider Kavanagh CM-S, Abby Reisman, Brad Fogo, Sarah McGrew, Peter Cipparone.
Teaching content in practice: Investigating rehearsals of social studies discussions. Teaching and
Teacher Education. 2019; 86. https://doi.org/10.1016/j.tate.2019.06.017 PMID: 31885415
57. Brint S, Cantwell AM, Saxena P. Disciplinary Categories, Majors, and Undergraduate Academic Experi-
ences: Rethinking Bok’s "Underachieving Colleges" Thesis. Research in Higher Education. 2012; 53
(1):1–25. https://doi.org/10.1007/s11162-011-9227-2
58. Hrastinski S. Introducing an informal synchronous medium in a distance learning course: how is partici-
pation affected? The Internet Higher Education. 2006; 9(2):117–131. https://doi.org/10.1016/j.iheduc.
2006.03.006
59. Huang Q. Comparing teacher’s roles of F2f learning and online learning in a blended English course.
Computer Assisted Language Learning. 2019; 32(3):190–209. https://doi.org/10.1080/09588221.2018.
1540434
60. Knowlton DS. A Theoretical Framework for the Online Classroom: A Defense and Delineation of a Stu-
dent-Centered Pedagogy. New Directions for Teaching and Learning. 2000; 2000(5–14). https://doi.
org/10.1002/tl.841
61. Duncan-Howell. Teachers making connections: Online communities as a source of professional learn-
ing. British Journal of Educational Technology. 2010; 41:324–340. https://doi.org/10.1111/j.1467-8535.
2009.00953.x
62. Partlow KM, Gibbs W. J. Indicators of constructivist principles in internet-based courses. Journal of
Computing in Higher Education. 2003; 14(2):68–97.
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 20 / 21
63. Smeby JC. Disciplinary differences in university teaching. Studies in higher education. 1996; 21(1):69–
79. https://doi.org/10.1080/03075079612331381467
64. Boys C, Brennan J., Henkel M., Kirkland J., Kogan M., Youl P. Higher Education and Preparation for
Work. Jessica Kingsley Publishers. 1988. https://doi.org/10.1080/03075079612331381467
65. Donald JG. Knowledge structures: Methods for exploring course content. Journal of Higher Education.
1983; 54:31–41.
66. Vo CZ M. H., and Diep A. N. Examining Blended Learning Implementation in Hard and Soft Sciences: A
Qualitative Analysis. International Journal of Research in Education and Science. 2020; 6(2):250–272.
67. Fang-fang G, Jing-huan S. Study on the Relationship between Curriculum-based Assessment and
Undergraduate Learning of Academic Disciplines. Higher education research. 2014; 35(5):63–70. cnki:
sun:high.0.2014-05-014
PLOS ONE
COVID-19’s impacts on the scope, effectiveness, and interaction characteristics of online learning
PLOS ONE | https://doi.org/10.1371/journal.pone.0273016 August 23, 2022 21 / 21
... During the Covid-19 pandemic, educators must ensure that learning activities continue, even though students are at home or in their respective places. An educator can use various online applications to support the achievement of learning objectives [3]. The application of learning with online applications is expected to help students learn optimally and effectively with a more interesting learning process even without having to do it face-to-face [4]. ...
Full-text available
Article
Understanding the material's content is one of the main competencies a teacher must possess, so it is necessary to diagnose the extent of understanding the level of understanding of science concepts for prospective teachers. This study aims to produce a valid and practical diagnostic assessment instrument assisted by live worksheets to measure the level of understanding of science concepts for elementary school teacher candidates. This research is development research using a 4-D model (define, design, develop and disseminate). In the define stage, science concepts that are difficult to understand for prospective elementary school teacher students have been identified, namely the material of temperature and heat, electricity, magnetic fields, and the solar system. The next stage is to design the right instrument to diagnose the level of understanding of science concepts for elementary school teacher candidates, namely a diagnostic instrument with three-level multiple choice questions integrated into the live worksheet application. Furthermore, in the development stage, the questions pay attention to the final abilities expected to be achieved in students based on the semester learning plans for special science education courses on materials identified at an early stage. The instrument that has been developed is then tested for validity and practicality. The validity test results obtained an average score of 3.8 which means the instrument is very valid, and the results of the practicality test show a practical instrument with an average score of 3.5. The final stage of this research is dissemination, namely the process of disseminating research and development products in the form of diagnostic assessment instruments assisted by live worksheets to measure the level of understanding of the science concepts of elementary school teacher candidates that have been developed. In conclusion, the live worksheet-assisted diagnostic assessment instrument developed is valid and practical for measuring the level of understanding of elementary school teacher candidates.
... The results of the present study are further strengthened with findings from an earlier study, which showed that all the changes made during the COVID-19 pandemic showed learning effectiveness [3]. Another study showed that the network scale of most courses did not change significantly, giving evidence that the COVID-19 pandemic did not notably change the scale of course interaction [45]. Interestingly, even before the COVID-19 pandemic, a research study showed that online or blended learning for teaching clinical skills is equally effective and similar to traditional means, and it may even benefit the teaching of clinical skills in undergraduate nursing curriculum [46]. ...
Full-text available
Article
In recent times, online teaching and assessment have provided a great opportunity to explore better methods in medical education. There are inconsistent views concerning the effectiveness of online assessment. Hence, the present study aimed to evaluate online teaching and assessment methods in relation to face-to-face methods by comparing students’ performances. The students’ assessment results in two basic anatomy courses, which are part of the Doctor of Medicine and Biomedical Sciences programs at Sultan Qaboos University, were analysed. We compared the students’ mean scores and coefficient of variance in the multiple-choice written exams and the objective structured practical exams during the spring semesters of 2019, 2020, and 2021, containing face-to-face teaching and exams, partial online teaching and online exams, and online teaching and both proctored online and face-to-face exams, respectively. The sudden transition to online teaching and assessment halfway through the semester resulted in higher means and a lower coefficient of variance among students’ scores in both theory and practical exams. However, when the fully adopted online method of teaching and assessment was employed, the mean scores decreased, and the coefficient of variance increased to figures close to those witnessed before the pandemic, when teaching and assessment were face-to-face. This trend applied to both the Doctor of Medicine and Biomedical Sciences programs’ anatomy courses. The results indicate that online assessment of theoretical and practical anatomical knowledge is comparable to that of face-to-face assessment. However, proper planning and preparedness are mandatory to achieve the desired outcomes.
Full-text available
Article
In the field of learning analytics, mining the regularities of social interaction and cognitive processing have drawn increasing attention. Nevertheless, in MOOCs, there is a lack of investigations on the combination of social and cognitive behavioral patterns. To fill in this gap, this study aimed to uncover the relationship between social interaction, cognitive processing, and learning achievements in a MOOC discussion forum. Specifically, we collected the 3925 participants’ forum data throughout 16 weeks. Social network analysis and epistemic network analysis were jointly adopted to investigate differences in social interaction, cognitive processing between two achievement groups, and the differences in cognitive processing networks between two types of communities. Finally, moderation analysis was employed to examine the moderating effect of community types between cognitive processing and learning achievements. Results indicated that: (1) the high- and low-achieving groups presented significant differences in terms of degree, betweenness, and eigenvector centrality; (2) the stronger cognitive connections were found within the high-achieving group and the instructor-led community; (3) the cognitive processing indicators including insight, discrepancy, and tentative were significantly negative predictors of learning achievements, whereas inhibition and exclusive were significantly positive predictors; (4) the community type moderated the relationship between cognitive processing and learning achievements.
Full-text available
Chapter
The COVID-19 outbreak in early 2020 brought online learning to the forefront of education. Scholars in China and abroad have conducted many studies on online learning during the pandemic, but only a few have performed quantitative comparative analysis on this topic during and before the outbreak. The current paper presents social network analysis of a course hosted on China’s MOOC platform “icourse163”. Specifically, this study aimed to uncover (1) variations in the scale of online learning amid COVID-19; (2) the characteristics of online learning interaction during COVID-19; and (3) teachers’/teaching assistants’ authority over online learning before, during, and after COVID-19. Results revealed that the scale of online learning interaction increased greatly during the pandemic but was not effectively maintained once the outbreak was under control. Online learning interaction became more frequent during the pandemic and more effective thereafter. The roles of teachers/teaching assistants in online learning during the pandemic were not as noteworthy as expected.
Full-text available
Article
The shift to distance teaching and learning during the COVID-19 pandemic brought about a real challenge for both instructors and students. To face these difficulties in teaching undergraduate Chemistry courses at the University of Santo Tomas, a blended learning strategy in the context of teaching and learning of Physical Chemistry 1 and Analytical Chemistry for Chemical Engineering students were employed. Here, we present an online strategy that facilitated the transition from traditional face-to-face learning to full online instruction. This is a five-component blended learning strategy referred to as Discover, Learn, Practice, Collaborate and Assess (DLPCA). In DLPCA, the asynchronous part of the teaching was achieved through broadcast of pre-recorded lecture videos on YouTube to allow students to study and progress with learning at their own pace. The synchronous part of the teaching was conducted using video conferencing platforms, such as Zoom or Google Meet. The DLPCA strategy was presented and discussed to the students prior to its implementation. The analysis of the teaching and learning experience based on three indicators (i) student’s learning experience, (ii) student’s academic performance and (iii) instructor observations showed that DLPCA had a positive impact on students and instructors. The identified challenges were stability of internet connection and instructor’s familiarity with readily available internet-based teaching tools, such as video conferencing software. Instructors must also find means to improve their interaction with students and maintain student interest and engagement during online classes. The survey also indicated that most of the students are satisfied with the DLCPA strategy. Hence, this strategy is considered a manageable and effective alternative that can be adapted to full online instruction to other undergraduate Chemistry lecture courses. Overall, the findings and insights in this study will add valuable resources for further hybrid instruction in the post-COVID-19 time in higher education.
Full-text available
Article
Although some studies have explored massive open online courses (MOOCs) discussion forums and MOOC online reviews separately, studies of both aspects are insufficient. Based on the theory of self-determination, this paper proposes research hypotheses that MOOC learning progress has a direct impact on MOOC online reviews and an indirect influence on MOOC online reviews through social interactions in discussion forums, as well. Coursera the largest MOOC platform, is selected as the empirical research object, and data from learners who participated in the MOOC discussion forum and provided MOOC online reviews from August 2016 to December 2019 are obtained from the most popular course, “Machine Learning”. After processing, data from 4376 learners are obtained. Then, according to research hypotheses, multi regression models are constructed accordingly. The results show that the length of MOOC online review text is affected by the MOOC learning progress, the number of discussion forum posts, the number of follow, the online review sentiment and MOOC rating. This study highlights the main factors that affect MOOC online reviews. As a result, some suggestions are put forward for the construction of MOOC.
Full-text available
Article
Background The COVID-19 pandemic has become a global health issue and has had a major impact on education. Consequently, half way through the second semester of the academic year 2019/2020, learning methods were delivered through distance learning (DL). We aimed to evaluate the student perspective of DL compared to classroom learning (CL) in the undergraduate dentistry study program at the Faculty of Dentistry Universitas Indonesia. Methods An online questionnaire was sent at the end of the semester. A total of 301 students participated in the study. Results Duration of study influenced student preference. Higher number of first-year students preferred DL compared to their seniors (p < 0.001). Students preferred CL for group discussion, as DL resulted in more difficult communication and gave less learning satisfaction. Only 44.2% students preferred DL over CL, although they agreed that DL gave a more efficient learning method (52.6%), it provided more time to study (87.9%) and to review study materials (87.3%). Challenges during DL included external factors such as unstable internet connection, extra financial burden for the internet quota and internal factors such as time management and difficulty to focus while learning online for a longer period of time. Conclusion Despite some challenges, dental students could adapt to the new learning methods of full DL and the majorities agreed blended learning that combined classroom and distance learning can be implemented henceforth. This current COVID-19 pandemic, changes not only the utilization of technology in education but the pedagogy strategies in the future.
Full-text available
Article
Due to COVID-19, higher education institutions transitioned to online learning. This study explored college students’ perceptions of their adoption, use, and acceptance of emergency online learning. The factors analyzed were attitude, affect, and motivation; perceived behavioral control (ease of use of technology, self-efficacy, and accessibility), and cognitive engagement. Quantitative and qualitative data were collected from 270 students. The findings present how motivation, self-efficacy, and use of technology play a significant role in the cognitive engagement and academic performance of students. Also, participants preferred face-to-face learning. This study presents suggestions on how to improve the acceptance of emergency online learning.
Full-text available
Article
The COVID19 pandemic has disrupted the normal functioning of various activities across the world, including learning and education. The shift towards online education during the pandemic of COVID19 has led many studies to focus on perceived learning outcomes and student satisfaction in this new learning environment. This study aims to examine the determinants resulting in students’ perceived learning outcomes and their influence on student satisfaction. The data was collected from undergraduate students in both South Korea and India to gain a cross-country study. The study found that the factors–interaction in the classroom, student motivation, course structure, instructor knowledge, and facilitation–are positively influencing students’ perceived learning outcome and student satisfaction. There is no significant difference in the students’ perceived learning outcome and student satisfaction in the two countries. The study will be helpful for the educationists and academics to identify the factors which will enhance student learning outcome and satisfaction level in online classes during the coronavirus pandemic.
Full-text available
Article
Educational institutes have closed worldwide due to the outbreak of COVID‐19. The research paper explores different instructional strategies. These instructional strategies can help in implementing online teaching in higher education. The case study develops various methods for online teaching without compromising student learning. These methods will help in designing successful online study sessions. The study formulated different notions for online education in developing countries. It includes maintaining slow voice and practicing vocal functions by teachers. Furthermore, sharing resources before the class will help in creating interactive online classes. These strategies play fundamental role in enhancing student learning. The study suggested that Higher Education Commission should make collaborations with telecommunication industries. This will help in overcoming Internet‐related issues. There are multiple factors that help in enhancing remote learning. These factors include getting feedback from students, offering flexible teaching and assessment policies. Recording online lectures and getting support from teaching assistants are also vital features. These instructional strategies will serve as an excellent tool in running online classes.
Full-text available
Article
Purpose This study aims to explore whether online learning has an effect on communication between instructors and students in a negative way, whether online learning affects students' productivity levels and to evaluate and suggest ways of improving effective online communication between instructors and students. Design/methodology/approach This study used is a quantitative research study which was conducted through a semi-structured online survey through a random sample technique. Findings Results revealed that the vast majority agree with the questions of the study. Students still prefer classroom classes over online classes due to many problems they face when taking online classes, such as lack of motivation, understanding of the material, decrease in communication levels between the students and their instructors and their feeling of isolation caused by online classes. Research limitations/implications This research studied the impact from students' perspective only as the sample was selected only from students. Originality/value This research reached the students’ point of view in a broader way which will help understanding the issues and provide effective solutions. This research suggested that instructors must communicate with their students and vice versa in more informal channels (instant messages online chat groups, audio calls, private video calls …) in parallel with the formal channels (online platforms, email …). Finally, instructors should encourage students to participate and study more by providing different kind of incentives.
Full-text available
Article
This study explores the perceptions of primary school teachers of online learning in a program developed in Indonesia called School from Home during the COVID-19 Pandemic. Data were collected through surveys and semi-structured interviews with 67 class teachers in primary schools. Data analysis used thematic analysis of qualitative data. The analysis results found four main themes, namely, instructional strategies, challenges, support, and motivation of teachers. This research contributes to the literature of online collaborative learning between teachers, parents, and schools that impact student success. Broadly, the success of online learning in Indonesia during the COVID-19 Pandemic was determined by the readiness of technology in line with the national humanist curriculum, support and collaboration from all stakeholders, including government, schools, teachers, parents and the community.